关键词: 天文交叉证认, HTM, HEALPix, PCAM并行程序设计模型, 大规模数据处理

Abstract: Cross-match is the kernel technology to realize multi-band data aggregation. It must be resolved by computer technologies since the astronomical data sets are very large. This paper issues a parallel crossmatch function using PCAM parallel programming model. Because the performance bottle-neck of cross-match exists on data I/O accessing, this function balanced the amount of data reading and computation by tuning partition granularity, removing blank areas, improving data load and computation flow, and etc. Experiments prove that it can speed up large-scale crossmatch greatly. On the other hand, the support of both of the two most famous astronomical data index functions HTM and HEALPix were considered, and the performance comparison between them were analyzed through experiments. The high-efficiency cross-match function in this paper offers a technology conference for the projects such as national astronomical database and virtual observatory.

Key words: astronomical cross-match, HTM, HEALPix, PCAM parallel programming model, massive data processing